Neonatal sepsis is a major cause of morbidity and mortality. The long term goal of this research is to improve outcomes of premature infants at risk of sepsis. The approach is based on the finding that clinical findings of neonatal sepsis are preceded by abnormal heart rate characteristics (HRC) of reduced variability and transient decelerations. Prior studies have led to a predictive algorithm for sepsis using continuous HRC monitoring that was developed at 1 center, and both internally and externally validated. Thus HRC monitoring is a potential adjunct to clinical care of premature infants that may improve their outcomes by allowing early diagnosis of neonatal sepsis. The Aim of this proposal is to test the hypothesis that HRC monitoring improves outcomes in NICU patients. The study design is a randomized clinical trial in which monitoring results are available for half the infants in 3 university tertiary care neonatal intensive care units. The main study group is very low birth weight infants, and the main outcome variable is ventilator-free days during the first 120 days of life. Secondary endpoints are length of stay, antibiotic usage, and in-hospital mortality, and we will study the supergroup of all NICU admissions. The site PIs are the division heads, and each has extensive experience in clinical trials. The project PI made the original observation of abnormal HRC prior to clinical signs of neonatal sepsis, and developed the validated clinical predictive tool. The Clinical Trials Office of the University of Virginia will provide secure and confidential data management. There is expert mathematical and biostatistical analysis. We feel that there is no more important problem in clinical neonatology, and that this approach represents a truly novel method to assist clinical neonatologists in early diagnosis of sepsis.